Image colourization 🖍🎨 for our project in the DTU Deep Learning course (02456)
- Data: use the places365 dataset (remove BW images)
- Make the baseline (GAN and L1-loss without transfer learning)
- Test difference between L1 and L2 loss on baseline model
- 2 backbones VGG19, Xception
- Quantitative evaluation (colourfulness, peak signal-to-noise ratio (PSNR))
- Qualitative human evaluation (by us) on 5 images each (discussion in report)
- Use image labels as additional conditional data and assess improvement
- Evaluate how image label data improved the model
$ are terminal commands
- open terminal in same folder as this project and type the following commands (you can paste them into the terminal with middle mouse click)
$ module load python3/3.9.6
$ module load cuda/11.3
$ python3 -m venv DeepLearning
$ source DeepLearning/bin/activate
$ pip3 install -r requirements.txt
Now everything should be setup. Then see the submit.sh
shell script for how it is activated. It should be run from the same path as the project.